Bayesian seasonal autoregressive integrated moving average: Modelling two decades of inflation dynamics in Nigeria

Abstract

There is a growing recognition of the unique characteristics of individual economies, particularly in developing countries. These peculiarities necessitate country-specific studies, as they provide insights tailored to the economic realities and challenges of the nation  under examination. This study was motivated by the need to explore these specific dynamics. This study employed Bayesian statistical  methods to model and forecast inflationary dynamics in Nigeria over two decades (January 2003 to September 2024). Adopting a  Bayesian Seasonal Autoregressive Integrated Moving Average, (SARIMA) framework, the analysis incorporates prior knowledge and  provides robust uncertainty quantification in parameter estimation and forecasting. Markov Chain Monte Carlo (MCMC) technique was  used to sample from posterior distributions, yielding insights into the persistence of inflation volatility and its implications for monetary  policy. This approach provides decision-makers with actionable forecasts and credible intervals, contributing to the literature on Bayesian  time series modeling in developing economies.